Add Docker Deployment

.devcontainer/devcontainer.json

@@ -6,7 +6,7 @@
     "codespaces": {
       "openFiles": [
         "README.md",
-        "app.py"
+        "conda_metadata_app/pages/main_page.py"
       ]
     },
     "vscode": {
@@ -19,7 +19,7 @@
   },
   "updateContentCommand": "[ -f packages.txt ] && sudo apt update && sudo apt upgrade -y && sudo xargs apt install -y <packages.txt; [ -f requirements.txt ] && pip3 install --user -r requirements.txt; pip3 install --user streamlit; echo '✅ Packages installed and Requirements met'",
   "postAttachCommand": {
-    "server": "streamlit run app.py --server.enableCORS false --server.enableXsrfProtection false"
+    "server": "streamlit run app_proxy.py --server.enableCORS false --server.enableXsrfProtection false"
   },
   "portsAttributes": {
     "8501": {

.dockerignore

@@ -0,0 +1,3 @@
+Dockerfile
+.pixi/
+.github/

.github/workflows/container.yml

@@ -0,0 +1,50 @@
+name: Container
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  build-docker:
+    runs-on: ubuntu-latest
+    name: Docker
+
+    permissions:
+      contents: read
+      packages: write
+
+    outputs:
+      digest: ${{ steps.build-and-push.outputs.digest }}
+
+    steps:
+      - name: Checkout branch
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938 # v4.2.0
+
+      - name: Login into GHCR
+        if: github.ref_name == 'main'
+        uses: docker/login-action@9780b0c442fbb1117ed29e0efdff1e18412f7567 # v3.3.0
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Docker meta
+        uses: docker/metadata-action@8e5442c4ef9f78752691e2d8f8d19755c6f78e81 # v5.5.1
+        id: meta
+        with:
+          images: |
+            ghcr.io/${{ github.repository }}
+          tags: | # no explicit versioning, always overwrite latest
+            type=raw,value=latest,enable=true
+      - name: Build and push
+        id: build-and-push
+        uses: docker/build-push-action@4f58ea79222b3b9dc2c8bbdd6debcef730109a75 # v6.9.0
+        with:
+          push: ${{ github.ref_name == 'main' }}
+          file: Dockerfile
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}

Dockerfile

@@ -0,0 +1,27 @@
+FROM ghcr.io/prefix-dev/pixi:0.31.0 AS build
+
+RUN apt update
+RUN apt install -y git
+
+COPY . /app
+WORKDIR /app
+
+RUN ./docker/install.sh
+
+FROM ubuntu:24.04 AS production
+
+COPY --from=build /app/.pixi/envs/default /app/.pixi/envs/default
+COPY --from=build /app/app_proxy.py /app
+COPY --from=build /app/app_config.toml /app
+COPY --from=build --chmod=0555 /entrypoint.sh /
+
+WORKDIR /app
+EXPOSE 8080
+ENTRYPOINT ["/entrypoint.sh"]
+CMD ["streamlit", \
+     "run", \
+     "--server.headless=true", \
+     "--global.developmentMode=false", \
+     "--browser.gatherUsageStats=false", \
+     "--server.port=8080", \
+     "app_proxy.py"]

README.md

@@ -8,10 +8,33 @@ A streamlit app to query metadata from conda packages.
 ![Main Screenshot](docs/assets/main_screenshot_light.png#gh-light-mode-only)
 
 ## Development
+Start by running `pixi run postinstall` to install the project locally.
+
 Use `pixi run dev` to run the app in development mode, and `pixi run deploy` to run a production version of the app.
 
 If you modified the configuration schema (see below), use `pixi run schema` to update the schema.
 
+### Dependencies
+Please note that this project defines dependencies in both a pixi project file (`pixi.toml`) and a `requirements.txt` file.
+The pixi project is used for local development and the Docker image, while the `requirements.txt` file is used for the
+public Streamlit cloud deployment.
+
 ## Custom Configuration
 
-Refer to the [Configuration Documentation](docs/configuration.md) for more information on how to customize the app.
+Refer to the [Configuration Documentation](docs/configuration.md) for more information on how to customize the app.
+
+## Docker Deployment
+A public Docker image of this app is available. To run the Docker app, execute the following command:
+
+```bash
+docker run -p 8080:8080 ghcr.io/Quansight-Labs/conda-metadata-app:latest
+```
+
+By default, the image uses the default configuration located at [app_config.toml](app_config.toml).
+
+To supply a custom configuration, mount a file to `/app/app_config.toml`:
+
+```bash
+docker run -p 8080:8080 -v /path/to/app_config.toml:/app/app_config.toml ghcr.io/Quansight-Labs/conda-metadata-app:latest
+```
+Note that if you use environment variables or secret files for credentials, you will need to set/mount those as well.

app_proxy.py

@@ -0,0 +1,8 @@
+"""
+This workaround is needed because streamlit does not support an app from a module.
+https://github.com/streamlit/streamlit/issues/662#issuecomment-553356419
+"""
+
+import runpy
+
+runpy.run_module("conda_metadata_app.app", run_name="__main__", alter_sys=True)

conda_metadata_app/app.py

@@ -0,0 +1,24 @@
+from pathlib import Path
+
+import streamlit as st
+
+"""
+If deploying a streamlit app as a Python module, we cannot use
+the automatic pages/ subpages. Instead, we need to define the
+pages manually.
+"""
+
+pages_dir = Path(__file__).parent / "pages"
+
+main_page = st.Page(
+    pages_dir / "main_page.py",
+    title="app",
+)
+
+search_by_file_path_page = st.Page(
+    pages_dir / "search_by_file_path_page.py",
+    title="Search By File Path",
+)
+
+pg = st.navigation([main_page, search_by_file_path_page])
+pg.run()

app.py → conda_metadata_app/pages/main_page.py

@@ -16,8 +16,9 @@
 from rattler.platform import PlatformLiteral
 from requests.auth import HTTPBasicAuth
 
-from app_config import AppConfig, Channel, PackageDiscoveryChoice, ArchSubdirDiscoveryChoice, ArchSubdirList, \
-    ArtifactDiscoveryChoice, MetadataRetrieval, Secret
+from conda_metadata_app.app_config import (AppConfig, Channel, PackageDiscoveryChoice, ArchSubdirDiscoveryChoice,
+                                           ArchSubdirList, ArtifactDiscoveryChoice, MetadataRetrieval, Secret)
+from conda_metadata_app.version_order import VersionOrder
 
 if not os.environ.get("CACHE_DIR"):
     from conda_oci_mirror import defaults
@@ -36,8 +37,6 @@
 from streamlit.logger import get_logger
 from xml.etree import ElementTree as ET
 
-from version_order import VersionOrder
-
 
 yaml = YAML(typ="safe")
 yaml.allow_duplicate_keys = True

pages/Search_by_file_path.py → conda_metadata_app/pages/search_by_file_path_page.py

@@ -5,7 +5,7 @@
 import streamlit as st
 from streamlit_searchbox import st_searchbox
 
-from app import app_config
+from conda_metadata_app.app_config import AppConfig
 
 
 @st.cache_resource(ttl="15m", max_entries=100)
@@ -49,7 +49,7 @@ def find_artifacts_by_path(path):
     initial_sidebar_state="collapsed",
 )
 
-if not app_config().enable_filepath_search:
+if not AppConfig().enable_filepath_search:
     st.error("File path search is disabled in the app configuration.")
     st.stop()
 

docker/install.sh

@@ -0,0 +1,10 @@
+#!/usr/bin/env bash
+
+set -eux
+
+pixi run -e build build-wheel
+pixi install -e default --locked
+pixi run -e default postinstall-production
+echo "#!/bin/sh" > /entrypoint.sh
+pixi shell-hook -e default -s bash >> /entrypoint.sh
+echo 'exec "$@"' >> /entrypoint.sh